Today on episode one of Dream Labs, we will be discussing a topic that I could not be more excited about how I made, yes, that's right. I'm talking about that little Diddy that you just heard when the show started. Goes a little bit like this.
That's right. No band, no orchestra, no singers. That sample that you just heard was generated completely by ai. Okay. Well, not completely by AI because. I also had a big part in the creative process. You see, SUNO, the leading app for AI generated music, has just dropped a new feature that I have been waiting for my entire life.
When you use this feature, you can upload an audio sample that you created yourself, and then give the AI prompt about what kind of song to generate with it. It lets you take a clip like this, which I hummed.
An output, something that sounds like this...
y'all, this is so great. It's like a mid journey moment for music, but actually it's much better because when you think about it, when you generate images on Mid Journey, you're still limited by what you can describe in text for what you generate. Or you could use a tool like CREA that's more advanced and actually lets you input images that the images that get output are based on.
But in order for that to work, you still need to have some basic drawing skills. And while you could say the same about this new way of making music with Suno, I personally think that more people are gonna have a little Diddy that they can hum out and then create something really cool with. Then are gonna have the tool set and skillset to actually be able to generate an image.
That's a good starting point for AI generated artwork. So what we're about to see is an explosion in musical creativity. People who never thought that they could have been musicians in the past are now just going to be able to take any idea that they can hum and turn it into a full fledged song. I don't know what this is gonna do for the music industry, but I'm very excited about what it'll mean for me personally.
'cause I have a lot of crazy song ideas and I want to express 'em. So let's hear another song that I just recently made using this technology.
Now look, if you could hear the input to this, you would think it was hilarious. It's absolutely ridiculous. Oh, you actually want me to play it? Okay. Okay, fine. Here it is.
So that input created this beat, which it didn't have lyrics. Then I downloaded this beat, wrapped over it and re-uploaded that to create the final product.
A piece that I still feel like I'm missing is being able to really hone in the voice that ends up on the final track, so that it sounds exactly the way that I want right now. I don't feel like I have the ability to control that voice to sound a little bit more like my own voice. Now, I don't want it to sound exactly like me.
But I do want to be able to create the specific voice that I'm trying to, and I don't feel like I have that ability yet. There's another piece of software that I think is much better for this. It's actually one of our portfolio companies, ACE Studio. Ace Studio allows you to create custom voices and generate lyrics in those voices in a way that Suno still really doesn't give you control over.
So if I wanted to create a fully finalized version of this song or others, what I probably would end up doing is generating the beat. And then creating the choruses using suno. But probably for those raps where I actually want them to sound more like myself, I'll end up using a studio to create a rap voice that more closely mirrors my own voice.
So here's an example of a song that I made using that type of workflow where I used my own voice to do the wraps and used a studio to create a different version of my voice for the chorus.
So look, while the sound of my voice on the track may. Not be as conventionally good as what you get with the Suno autogenerated voice. I still feel like that's a much more accurate expression of my own musical inspiration than what Suno generated, and I think that that's gonna become more and more important as we move forward.
Because as you can see, it's now easy to generate any kind of typical music with Suno and other AI audio tools. And so what's gonna really set pieces apart, I think, is going to be how much self-expression there is in that piece. And I think we're gonna see a world where more and more people are generating what feels true to them, and more people are expressing themselves musically than ever before.
And in ways you probably won't expect, maybe a little bit like this.
So this one I generated the beat using Wandera, which is another one of our portfolio companies. That's. Portfolio of music that their model is trained on is all completely copyright free. The thing about generating music with Suno is that they've actually used a bunch of copyrighted music to train their models.
So the music that you generated with it, you can't really use as professional music that you would publish. Wandera on the other hand, has a completely clean data set, but it's all from karaoke music. So honestly, the outputs still aren't quite as good. But there's still interesting things that you can do with it.
For example, in that song that I just played for you, which was admittedly quite bizarre, I kept pushing the Wandera model to generate weirder and weirder songs and patterns of notes that had never been played before in the history of music. So let's take a look at another song that I made this time. I wanted to create a dance track that I could play on the playa at Burning Man.
So here's the little Diddy that I sang into my microphone to get it started.
Beautiful. I know, but check out what it generated.
Pretty epic, right? The inspiration for the song actually came from another one of our portfolio companies, generation lab. They're working on a longevity diagnostic that has given me the belief that we are going to have the. Choice of how long we want to live when we are older. The only limitation might end up being the size of our pocketbooks.
But here's the thing, this first version of the song that I generated, I didn't really like the style of the electronic music. It's not very modern. It's not the kind of melodic house that you hear on the playa these days. So I did another remix and got this result.
Now that has Mayan Warrior Sunset set, written all over it. It. And speaking of Burning Man, I wanted to create some custom theme music for our camp. It's called Celestial Tea Garden, and it's gonna be located at six 15 and D. So if you'll be there at the burn this year, definitely come see us. And if you do, you might just hear our special theme song.
And now I have to give a special shout out because that camp was started by two of our other portfolio founders at HF zero o, Shannon and Jesse from Open Home. Open Home is a smart speaker operating system that you can also use as a developer to develop. Smart voice apps and I wouldn't be surprised if in the near future there will be an open home app where you can just hum the song that you want to hear and get back the fully rendered version of it right there on your smart speaker.
So look, you might've thought that you knew a thing or two about music production.
And I hope you enjoyed this episode of Dream Labs.